NEURA Robotics and TUM open RoboGym in Munich, unveil 4NE1 Gen 4 humanoid

A training center where companies teach robots their tasks

The RoboGym was built in partnership with TUM's Munich Institute of Robotics and Machine Intelligence (MIRMI). According to NEURA, it brings together research, robot training and industrial application. Companies contribute concrete tasks and their process knowledge, while researchers and development teams study how robots can learn these tasks faster and perform them more reliably.

Partners have been training workflows at the site since September. Companies from more than ten industries are involved, including automotive, pharmaceuticals, medical technology, packaging, logistics, agriculture and care. NEURA says demand is high and that capacity in Munich is already in strong demand. Tasks range from handling cardboard boxes and filling bottles to loading and unloading wooden components.

Companies can test applications, identify training needs and assess how suitable different robot systems are before expanding their use in their own operations. The available form factors include humanoid robots, mobile assistance robots such as MiPA, and cognitive robot arms.

Three ways to generate physical training data

NEURA argues that Physical AI needs data on how objects, materials and forces behave in the real world. The RoboGym combines three approaches to produce it:

• Real-world robot data: different robot systems perform tasks, and their sensors capture movements, contacts and outcomes.
• Human demonstrations: people show workflows, with sensor-equipped data suits, motion-capture systems and teleoperation turning movements into usable training data.
• Simulation and Data Amplifier: virtual environments add variations and situations, so skills can be trained under changed conditions and then validated on real robots.

The Neuraverse connects this training infrastructure with NEURA's software and AI tools. For the joint research by TUM and NEURA, the training data is structured and processed for developing and studying Physical AI models.

4NE1 Gen 4: lighter, slimmer, built for series production

According to NEURA, 4NE1 Gen 4 ("For Anyone") is designed for serial production and is 20 percent lighter than its predecessor. It has a slimmer body and a black textile outer skin, which the company says makes the robot appear more approachable and offers potential for simple customization. The robot is already in operation at the TUM RoboGym and the NEURA Gyms.

NEURA also points to more precise handling. Advanced hands are meant to enable more accurate grips, and so-called Smart Limbs have reflexes built directly into each limb, so each can respond independently. The company says this lets the robot react instantly to its environment and move more fluidly.

The robot runs on Neuron AI, which NEURA describes as its Perception-Language-Action model for Physical AI. Besides vision, touch, force, sound and the robot's own state feed into the model. NEURA cites the example of recognizing whether a fastener has clicked into place. Through the Neuraverse, skills learned by one robot become available to other connected robots. 4NE1 Gen 4 is also appearing at the NVIDIA GTC conference in Berlin.

Funding, partners and the planned Gym network

The joint initiative is initially set up for three years. The partners are investing around EUR 17 million in total, of which NEURA contributes approximately EUR 11 million. NEURA provides robotics, technical infrastructure, computing power and operating staff, among other things, while TUM supplies space, scientific expertise and resources for research and operations.

"The TUM RoboGym robotics center is unique in Germany and Europe," said Prof. Thomas Hofmann, President of TUM. David Reger, Founder and CEO of NEURA Robotics, said the company wants to build "a powerhouse for physical capabilities" with TUM and that the Mittelstand in particular should be able to help shape this future.

Munich is the start of a planned international network of NEURA Gyms. Further locations are being established in Europe, the United States and China, and another Gym in Germany is being built in cooperation with RWTH Aachen.

What it means for industry users

For companies, the RoboGym offers a way to try robots on their own tasks before committing to wider deployment. Firms can test applications and compare robot systems in a shared environment that also includes academia and startups. Because skills learned in one Gym are meant to be shared across connected robots, NEURA says 4NE1 Gen 4 benefits from what is learned across the growing network. The source gives no pricing or delivery dates for the robot, which is described as being prepared for series production.